Velodyne pcl

[spuetz]

Regrading #287 The velodyne_pcl ROS package restores the conversion file point_types.h.
The README explains how to use it.

The following has been done:

• Fixed assignment of the time value in the organized point cloud container
• Added velodyne_pcl ROS package with the point_types.h header file.
• Added a RAEDME.md file to explain how to convert velodyne ROS point clouds to PCL
• Renamed both containers to cover the previously added time property.

[spuetz]

@JWhitleyWork Could you test the conversion? Cheers.

[JWhitleyWork]

@spuetz Just two requests if you could:

1. Please add a nodelet which simply does the conversion and a launch file that runs the nodelet with parameters for a nodelet manager and input and output topics.
2. Please add roslint to the package and the associated test.

[JWhitleyWork]

Thanks again for working on this.

[spuetz]

The conversion is also automatically done by the subscriber and the publisher. See http://wiki.ros.org/pcl/Overview#Publishing_and_subscribing_to_point_cloud_messages
I think it would make sense to just provide a test where we know which data is in the could and then compare it after conversion.
Would be good if someone could write the test, I don't have much time... and I think you are more into the testing of your laser scan data with the pcap data.

[JWhitleyWork]

Thanks for the clarification, @spuetz.

[JWhitleyWork]

@spuetz I'll add roslint to the package. However, I've also just realized that this MR breaks API in another way by doing away with the PointCloudXYZIR type. While I agree that the type name should match the contents, this name change breaks existing use of the type by many, many examples on Github. This change can be introduced in Noetic, but not in the Melodic LTS. Please revert this commit.

[spuetz]

Someone introduced the time field in that container, instead of creating a new one. That mainly caused the problem. Cleanest solution would be to add another container without the time field and to add the corresponding point type here.

These depending packages, you mentioned, they also need to change the include file to velodyne_pcl/point_types.h. Additionally they need to install the velodyne_pcl package.

[spuetz]

@JWhitleyWork Any news here? See my comment above.

[JWhitleyWork]

@spuetz Sorry I haven't responded lately. Yes, I understand that users will need to change the include path and depend on velodyne_pcl. Since these changes are necessary already, changing the point type doesn't seem like that big of a deal.

I'm creating a PR to your branch right now to add roslint.

[JWhitleyWork]

PR Created for roslint: spuetz#1

[YoshuaNava]

@JWhitleyWork @spuetz I'm looking forward to this PR being merged and would be glad to help test if needed. I have a VLP-16 that I can use for tests.

[JWhitleyWork]

@YoshuaNava I would really appreciate that! Could you possibly test this prior to merge? If you clone from @spuetz fork of the repository and check out his velodyne_pcl branch, you should be able to build and test it. Please let me know how it goes and I'll merge it immediately after!

[YoshuaNava]

@JWhitleyWork Can you send me a short list of the functionality I should text, and what is expected from each test?

[JWhitleyWork]

See comments.

[JWhitleyWork]

Last item: can you rebase on master so it uses Noetic CI?